A test of Harrington's experimental model of ethnic bias in testing applied to a measure of emotional functioning in adolescents /
Gordon Harrington, an experimental psychobiologist, performed a series of animal studies in the 1960s and 1970s that resulted in his formation of a hypothesis about the inevitability of ethnic bias in tests constructed using traditional procedures. Specifically, he hypothesized that the ethnic grou...
| Main Author: | |
|---|---|
| Format: | Thesis Book |
| Language: | English |
| Published: |
[Place of publication not identified] :
[publisher not identified] ;
1995.
|
| Subjects: | |
| Online Access: | http://proxy.library.tamu.edu/login?url=http://proquest.umi.com/pqdweb?did=741213241&sid=1&Fmt=2&clientId=2945&RQT=309&VName=PQD |
| Summary: | Gordon Harrington, an experimental psychobiologist, performed a series of animal studies in the 1960s and 1970s that resulted in his formation of a hypothesis about the inevitability of ethnic bias in tests constructed using traditional procedures. Specifically, he hypothesized that the ethnic group which comprised the majority of the test construction sample would influence item selection procedures to such an extent that the resulting test would be biased in their favor, and would be biased against any other ethnic groups. He further asserted that the results of his animal experiments were directly applicable to tests constructed for humans, as it was the test construction procedures themselves that were being studied, and about which his conclusions were drawn. This dissertation put his hypothesis to the test for a measure of emotional functioning in children using the standardization form of the Behavioral Assessment System for Children (BASC), Self-Report of Personality (Adolescent form) (SRP-A). Traditional item selection techniques were applied to items comprising the 14 subscales found on the standardization form of the BASC SRP-A, and new subscales were created for the black, Hispanic, and non-Hispanic white populations. Two experiments were performed in which 100% ethnic representation and variable ethnic representation comprised the test construction samples. Subsequent analyses were undertaken in a search for statistically and practically significant levels of differential functioning and diagnosis of ethnic groups, and included means, standard deviations, Tukey's HSD test, computation of effect sizes, reliability coefficients (Cronbach's alpha), item overlap/non-overlap, chi-square tests of independence, and ANOVAS. None of the analyses resulted in statistically or practically significant differences in performance or diagnosis among ethnic groups that would have been expected were Harrington's hypothesis true for this case. |
|---|---|
| Item Description: | Vita. "Major Subject: School Psychology". |
| Physical Description: | ix, 143 leaves ; 28 cm. Issued also on microfiche from University Microfilms Inc. |
| Bibliography: | Includes bibliographical references. |